Reconsidering Deception in Social Robotics: The Role of Human Vulnerability

Auteurs

Carli R., Najjar A.

Référence

Proceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI 2023, vol. 37, pp. 16174-16175, 2023

Description

The literature on deception in human-robot interaction (henceforth HRI) could be divided between: (i) those who consider it essential to maximise users' end utility and robotic performance; (ii) those who consider it unethical, because it is potentially dangerous for individuals' psychological integrity. However, it has now been proven that humans are naturally prone to anthropomorphism and emotional attachment to inanimate objects. Consequently, despite ethical concerns, the argument for the total elimination of deception could reveal to be a pointless exercise. Rather, it is suggested here to conceive deception in HRI as a dynamic to be modulated and graded, in order to both promote innovation and protect fundamental human rights. To this end, the concept of vulnerability could serve as an objective balancing criterion.